They started by mixing sour plum vodka at home, now they produce 1,800 bottles every month

At every party, people would pull Alexander Cheong aside and ask the same question: can I get a bottle of that?

The “that” was his homemade Sour Plum Vodka. Sick of alcohol brands that felt serious and disconnected from the actual experience of drinking, Alexander started mixing his own at social gatherings, rooted in the Southeast Asian flavours he grew up with.

“Alcohol is a social lubricant,” he said. “It’s about making the worries and stress clear up. We want to embody not taking anything too seriously.”

That philosophy became the foundation of Clumzy—a spirits brand he built with two friends, Kenneth Tan and Daniel Lim, on one simple idea: bottling Southeast Asian flavours with a bit of a kick. Five years later, with just three flavours and no outside investment, the brand has crossed S$1 million in cumulative revenue.

We spoke to the founders to find out how a kitchen experiment became a million-dollar business.

Starting a spirits brand from scratch

The origins of Clumzy trace back to 34-year-old Alexander’s natural flair for mixology. Always the life of the party, he rarely enjoyed what he called “cold, hard, and serious” alcohol. 

At social gatherings, to save money, he would mix his own cocktails, making his own flavourings from whatever he could imagine. Eventually, Alexander became known for his experimental jungle juices and punches, but one creation in particular stood out: a Sour Plum Vodka that became an instant crowd favourite at every party.

When COVID-19 hit, and social gatherings were restricted, demand for Alexander’s creations didn’t disappear—instead, it intensified. Friends and even friends of friends wouldn’t stop asking him to bottle his drinks for family occasions and casual nights in.

So, Alexander roped in his friend Kenneth to launch Clumzy in early 2021, taking orders via Instagram DMs only. Without any real push, word of mouth spread faster than they could keep up with.

But there was one clear problem: branding.

The product started with a simple label and packaging that wasn’t particularly eye-catching. About a month in, they brought in Daniel, an experienced marketer and close friend, as the third co-founder to help turn the product into a proper brand.

Daniel’s entry laid down Clumzy’s branding foundations. He redesigned the labels and packaging—joking that the original bottle looked like a “medicine bottle”—along with new photography, the website, and the e-commerce platform, shaping the brand identity Clumzy is known for today.

From there, while Daniel handled brand and marketing, Alexander and Kenneth focused on the business, trade relationships, and operations.

“If it didn’t work out, we had nothing to fall back on”

Clumzy started out with one sole offering: the Sour Plum Vodka. The product is marketed for its versatility—being great for shots, served on the rocks or as a cocktail. Each bottle retails at S$58 with an alcohol by volume of under 20%.

In the beginning, the spirit was made by the trio in Kenneth’s kitchen, which was stocked with giant Cambro containers and bottles. Supplies were stored in a small rented warehouse at S$200 per month, meaning they had to physically lug stock back and forth for every production run to Kenneth’s home.

At the time, they were making around 180 bottles a month—a capacity they quickly outgrew as demand surged beyond what a home setup could sustain.

Within just a few months, Clumzy became a legitimate side hustle generating real extra income on top of their day jobs. That created a genuine decision point for the co-founders: were they to stay comfortable with some pocket money, or risk everything to grow it into a real business?

Operating from a home kitchen came with clear limitations. They could only sell directly to consumers, with B2B opportunities completely off the table without a licensed commercial facility.

But upgrading wasn’t a small leap. Setting up a proper production space required tens of thousands of dollars—essentially their entire annual revenue at the time.

If it didn’t work out, we had nothing to fall back on.

Daniel Lim, co-founder of Clumzy

Still, they chose to take the risk. The trio secured a liquor licence and set up a dedicated production facility, moving from manual preparation to automated mixing and bottling processes.

They hosted pop up booths almost every weekend

From the start, the founders have been prudent with their spending, purchasing only what was necessary. They bootstrapped the venture with “a couple of thousand dollars,” which generated revenue and paid for itself, breaking even within a year.

Early on, they also committed to building Clumzy through direct consumer engagement.

They signed up for pop-ups and hosted booths “almost every weekend,” becoming familiar faces at curated events like ArtBox 2023, Boutiques Singapore 2024, and Christmas Atelier 2025, amongst smaller pop-ups at bars and cafes since 2021.

“Demand has been very strong,” Daniel noted. “We sold out on the first day of our first Boutiques Singapore back in 2022, which lasted two weeks. We made in one day what we normally would make over three to four days at other events.”

As Clumzy’s presence at weekend pop-ups grew, restaurants and eateries began taking notice, often after seeing strong customer demand at events or encountering the brand through word of mouth and social media.

At the same time, another key growth driver was how the founders expanded the ways customers could experience the product. They diversified into events, offering on-tap services for weddings and house parties, and experimented with more flexible formats, including slushie versions of their drinks at events.

“Diversifying into slushies made sense for us to create an option to make drinks that are fun for people who may not want to drink alcohol that tastes like alcohol,” Daniel shared.

He added that 65% of their customers are women, reflecting how Clumzy has resonated with a demographic that traditional spirits brands have historically underserved.

Hitting the S$1M milestone

Today, Clumzy produces around 1,800 bottles of spirits each month and has expanded to a team of eight. Its offerings are stocked at 11 retail partners, including Pat’s Music Pub and The Liquor Shop.

The business model currently sees revenue split roughly into 70% B2C and 30% B2B. After four years of operation, Clumzy crossed S$1 million in cumulative revenue in 2025, a significant milestone for a bootstrapped local spirits brand.

The brand has also since expanded its product line to three spirits, with customer feedback playing a crucial role in shaping its development.

While the Sour Plum Vodka developed a devoted following, it became clear that the polarising flavour was an acquired taste—some loved it instantly, while others needed time to warm up to it. This insight led to the development of the brand’s second flavour: the Chrysanthemum Lychee Gin, designed for those who find the sour plum variety too intense.

After persistent calls from customers for a third offering, the trio released Coconut Pandan Rum recently, a creation they felt was important to Clumzy’s identity as a business inspired by Southeast Asian flavours.

Protecting what they’ve built

Around the time Clumzy turned one, several parties approached the founders with offers: investment for equity stakes, rebranding under a larger umbrella, or outright acquisition.

At the time, all three founders still had day jobs. These offers initially felt genuinely attractive—a chance to take the brand one step up without carrying all the risk alone.

But Alex had the foresight to see what they’d be giving up: undervaluing everything that they stood for as a small founder-led brand. As such, they turned the offers down. 

“In hindsight, some of those offers really shortchanged us,” Alex said. “I’m glad we trusted our gut in those decisions, and I’m glad we saw it all pay off eventually through our hard work.”

Today, Clumzy is run by Alexander and Daniel, with Kenneth having taken a backseat in 2024.

The next chapter for the brand involves significantly expanded distribution. After months of negotiations with NTUC FairPrice, the brand’s spirits are set to hit its supermarket shelves soon, marking its entry into mainstream retail.

Encouraged by strong demand in Singapore, international expansion is also on the horizon. The founders plan to bring Clumzy to Thailand and Australia, driven by interest from Singaporeans abroad who have discovered the brand and want access in their new home countries.

That said, the team has also observed a broader shift in drinking culture, with more consumers becoming intentional about their alcohol consumption. While nightlife has seen a decline, overall alcohol consumption remains relatively steady, as more people drink at home or during daytime social occasions.

Daniel noted that it’s all about a healthier relationship with socialising.

“People aren’t drinking less. They’re bored with sameness. Alcohol offerings have felt largely unchanged for decades. What people are genuinely hungry for is novelty and a sense of connection to something familiar,” Daniel explained.

That sense of novelty and familiarity is what Clumzy aims to deliver.

• Learn more about Clumzy here.
• Read more articles we’ve written on Singaporean businesses here.

The Bremen startup’s platform deploys teams of AI agents that autonomously execute engineering tasks across more than 75 existing tools, without replacing any of them. Revaia led the Series B; Capgemini joined through its ISAI Cap Venture vehicle. All Series A investors returned.

Synera, the Bremen-based agentic AI platform for industrial engineering, has raised $40 million (approximately €35 million) in a Series B round led by Revaia, with participation from Capgemini through ISAI Cap Venture.

All of the company’s existing Series A investors returned, including UVC Partners with a substantial commitment from its growth fund, BMW iVentures, Cherry Ventures, Venture Stars, and Spark Capital.

The round is intended to accelerate Synera’s expansion in the US and internationally, building on existing deployments at NASA, BMW, Airbus, Volvo Trucks, and Hyundai.

Synera was founded in 2018 in Bremen by Dr. Moritz Maier, Sebastian Möller-Lafore, and Daniel Siegel, a team that had been working together since 2006, initially under the name ELISE (Evolutionary Lightweight Structure Engineering), before rebranding in 2022 to reflect the company’s expanded scope.

The platform connects more than 75 existing engineering tools, including software from Altair, Autodesk, Hexagon, PTC, and Siemens, into a unified orchestration layer, allowing AI agents to execute complex engineering tasks autonomously across design, simulation, optimisation, costing, and reporting without requiring companies to replace their existing infrastructure.

The platform is deployed on-premises, keeping engineering intellectual property and sensitive data within customers’ own environments. Synera has also established a US presence in Boston, Massachusetts.

The company describes its approach as deploying a virtual engineering team: agents that don’t merely assist but autonomously execute, running iterative simulations, generating reports, responding to RFQs, and progressing through approval workflows without human intervention at each step.

The platform has been internally described as “JARVIS for engineers.” Quantified outcomes cited by Synera and independently validated by Frost & Sullivan in a 2025 analysis include a 95% reduction in finite element simulation time at engineering consultancy EDAG, and a 30% weight reduction in 3D-printed robot gripper designs at BMW’s Additive Manufacturing Campus.

NASA has deployed multiple Synera agents to transform requirements into validated part designs, completing hundreds of design iterations in an hour.

The investment context is a structural mismatch between AI investment and manufacturing deployment. Gartner’s 2025 CIO survey found that 86% of manufacturing respondents plan to increase generative AI investment in 2026 and 97% expect to have deployed it by 2028, yet only 41% of AI and generative AI prototypes currently reach production, according to Gartner’s 2024 AI Mandates for the Enterprise survey.

Synera’s proposition is that the gap exists because most AI tools treat engineering as a chat interface problem rather than an infrastructure problem: the agents need to connect to the actual tools where the work happens, not sit alongside them.

The company has also been recognised by Frost & Sullivan with its 2025 Global AI Agents for Engineering Transformational Innovation Leadership award.

The Series A, raised in September 2022, was $14.8 million, led by Spark Capital with BMW iVentures, Cherry Ventures, UVC Partners, and Venture Stars participating. The Series B brings total funding to approximately $58 million.

Capgemini’s entry through ISAI Cap Venture is strategically notable: Capgemini is one of the world’s largest IT services firms and a significant engineering services provider to the automotive and aerospace sectors Synera targets, making it both an investor and a potential channel partner.

Pillar, a platform that helps commodity-driven businesses (like those in metals, food, and airline companies) manage financial risk, announced Tuesday a $20 million seed round led by Andreessen Horowitz. 

Others in the seed round include Crucible Capital, Gallery Ventures, and Uber CEO Dara Khosrowshahi. The company has raised $23 million to date.

Pillar, founded in 2023, automates hedging processes for such businesses. Hedging is when a company places a trade that can offset or cancel out losses from other priced trades. Geopolitics has not been kind to the commodities market, which has seen much volatility in the past year. 

Harsha Ramesh, the company’s co-founder and CEO (founded alongside Chinmay Deshpande, the company’s CTO), said the company uses AI to ingest and parse data from client contracts, cash flows, inventories, ERP software, spreadsheets, and even WhatsApp messages to “continuously analyze exposure across commodities, FX, and freight.”

It can then build and manage a hedge portfolio for its clients, and adjust positions automatically based on “market conditions, volatility, and the client’s risk tolerance,” Ramesh continued. The platform executes trades and continuously monitors risk and exposure, turning hedging from a “static, periodic decision to a continuous, autonomous system,” Ramesh said.

Pillar’s clients include Shibuya Sakura Industries, a trading firm that buys and sells commodities like metals; the recyclable materials company Sigma Recycling; and United Metals Solution Group, which also recycles and trades metals. 

Ramesh was once a macro trader, managing large derivative trading books and working with some of the largest companies in the world as they sought to hedge foreign exchange rates and interest rate exposure, he said. “I also spent time at a medium-sized physical business in import-export,” he recalled. 

“What stood out was that sophisticated institutions had access to tools, infrastructure, and talent, while the actual producers, importers, and manufacturers driving global trade had little to no access to this,” he said. “Risk management was treated as a luxury, despite being essential.” 

Pillar hopes to offer sophisticated, institutional-grade tools to small and medium-sized enterprises. “Our goal is to make hedging as accessible and ubiquitous as payments or accounting software,” he said.

Others in this business include the legacy desks at big banks and the commodity risk platforms like Topaz and RadarRadar.

Ramesh said humans are still in the loop in some ways at Pillar, handling “approvals, oversight, and strategic decisions.” Humans also help with more “complex situations” — like large transactions, where a human team will mix their judgment with the machine’s execution. 

The software industry is racing to write code with artificial intelligence. It is struggling, badly, to make sure that code holds up once it ships.

A survey of 200 senior site-reliability and DevOps leaders at large enterprises across the United States, United Kingdom, and European Union paints a stark picture of the hidden costs embedded in the AI coding boom. According to Lightrun’s 2026 State of AI-Powered Engineering Report, shared exclusively with VentureBeat ahead of its public release, 43% of AI-generated code changes require manual debugging in production environments even after passing quality assurance and staging tests. Not a single respondent said their organization could verify an AI-suggested fix with just one redeploy cycle; 88% reported needing two to three cycles, while 11% required four to six.

The findings land at a moment when AI-generated code is proliferating across global enterprises at a breathtaking pace. Both Microsoft CEO Satya Nadella and Google CEO Sundar Pichai have claimed that around a quarter of their companies’ code is now AI-generated. The AIOps market — the ecosystem of platforms and services designed to manage and monitor these AI-driven operations — stands at $18.95 billion in 2026 and is projected to reach $37.79 billion by 2031.

Yet the report suggests the infrastructure meant to catch AI-generated mistakes is badly lagging behind AI’s capacity to produce them.

“The 0% figure signals that engineering is hitting a trust wall with AI adoption,” said Or Maimon, Lightrun’s chief business officer, referring to the survey’s finding that zero percent of engineering leaders described themselves as “very confident” that AI-generated code will behave correctly once deployed. “While the industry’s emphasis on increased productivity has made AI a necessity, we are seeing a direct negative impact. As AI-generated code enters the system, it doesn’t just increase volume; it slows down the entire deployment pipeline.”

Amazon’s March outages showed what happens when AI-generated code ships without safeguards

The dangers are no longer theoretical. In early March 2026, Amazon suffered a series of high-profile outages that underscored exactly the kind of failure pattern the Lightrun survey describes. On March 2, Amazon.com experienced a disruption lasting nearly six hours, resulting in 120,000 lost orders and 1.6 million website errors. Three days later, on March 5, a more severe outage hit the storefront — lasting six hours and causing a 99% drop in U.S. order volume, with approximately 6.3 million lost orders. Both incidents were traced to AI-assisted code changes deployed to production without proper approval.

The fallout was swift. Amazon launched a 90-day code safety reset across 335 critical systems, and AI-assisted code changes must now be approved by senior engineers before they are deployed.

Maimon pointed directly to the Amazon episodes. “This uncertainty isn’t based on a hypothesis,” he said. “We just need to look back to the start of March, when Amazon.com in North America went down due to an AI-assisted change being implemented without established safeguards.”

The Amazon incidents illustrate the central tension the Lightrun report quantifies in survey data: AI tools can produce code at unprecedented speed, but the systems designed to validate, monitor, and trust that code in live environments have not kept pace. Google’s own 2025 DORA report corroborates this dynamic, finding that AI adoption correlates with an increase in code instability, and that 30% of developers report little or no trust in AI-generated code.

Maimon cited that research directly: “Google’s 2025 DORA report found that AI adoption correlates with an almost 10% increase in code instability. Our validation processes were built for the scale of human engineering, but today, engineers have become auditors for massive volumes of unfamiliar code.”

Developers are losing two days a week to debugging AI-generated code they didn’t write

One of the report’s most striking findings is the scale of human capital being consumed by AI-related verification work. Developers now spend an average of 38% of their work week — roughly two full days — on debugging, verification, and environment-specific troubleshooting, according to the survey. For 88% of the companies polled, this “reliability tax” consumes between 26% and 50% of their developers’ weekly capacity.

This is not the productivity dividend that enterprise leaders expected when they invested in AI coding assistants. Instead, the engineering bottleneck has simply migrated. Code gets written faster, but it takes far longer to confirm that it works.

“In some senses, AI has made the debugging problem worse,” Maimon said. “The volume of change is overwhelming human validation, while the generated code itself frequently does not behave as expected when deployed in Production. AI coding agents cannot see how their code behaves in running environments.”

The redeploy problem compounds the time drain. Every surveyed organization requires multiple deployment cycles to verify a single AI-suggested fix — and according to Google’s 2025 DORA report, a single redeploy cycle takes a day to one week on average. In regulated industries such as healthcare and finance, deployment windows are often narrow, governed by mandated code freezes and strict change-management protocols. Requiring three or more cycles to validate a single AI fix can push resolution timelines from days to weeks.

Maimon rejected the idea that these multiple cycles represent prudent engineering discipline. “This is not discipline, but an expensive bottleneck and a symptom of the fact that AI-generated fixes are often unreliable,” he said. “If we can move from three cycles to one, we reclaim a massive portion of that 38% lost engineering capacity.”

AI monitoring tools can’t see what’s happening inside running applications — and that’s the real problem

If the productivity drain is the most visible cost, the Lightrun report argues the deeper structural problem is what it calls “the runtime visibility gap” — the inability of AI tools and existing monitoring systems to observe what is actually happening inside running applications.

Sixty percent of the survey’s respondents identified a lack of visibility into live system behavior as the primary bottleneck in resolving production incidents. In 44% of cases where AI SRE or application performance monitoring tools attempted to investigate production issues, they failed because the necessary execution-level data — variable states, memory usage, request flow — had never been captured in the first place.

The report paints a picture of AI tools operating essentially blind in the environments that matter most. Ninety-seven percent of engineering leaders said their AI SRE agents operate without significant visibility into what is actually happening in production. Approximately half of all companies (49%) reported their AI agents have only limited visibility into live execution states. Only 1% reported extensive visibility, and not a single respondent claimed full visibility.

This is the gap that turns a minor software bug into a costly outage. When an AI-suggested fix fails in production — as 43% of them do — engineers cannot rely on their AI tools to diagnose the problem, because those tools cannot observe the code’s real-time behavior. Instead, teams fall back on what the report calls “tribal knowledge”: the institutional memory of senior engineers who have seen similar problems before and can intuit the root cause from experience rather than data. The survey found that 54% of resolutions to high-severity incidents rely on tribal knowledge rather than diagnostic evidence from AI SREs or APMs.

In finance, 74% of engineering teams trust human intuition over AI diagnostics during serious incidents

The trust deficit plays out with particular intensity in the finance sector. In an industry where a single application error can cascade into millions of dollars in losses per minute, the survey found that 74% of financial-services engineering teams rely on tribal knowledge over automated diagnostic data during serious incidents — far higher than the 44% figure in the technology sector.

“Finance is a heavily regulated, high-stakes environment where a single application error can cost millions of dollars per minute,” Maimon said. “The data shows that these teams simply do not trust AI not to make a dangerous mistake in their Production environments. This is a rational response to tool failure.”

The distrust extends beyond finance. Perhaps the most telling data point in the entire report is that not a single organization surveyed — across any industry — has moved its AI SRE tools into actual production workflows. Ninety percent remain in experimental or pilot mode. The remaining 10% evaluated AI SRE tools and chose not to adopt them at all. This represents an extraordinary gap between market enthusiasm and operational reality: enterprises are spending aggressively on AI for IT operations, but the tools they are buying remain quarantined from the environments where they would deliver the most value.

Maimon described this as one of the report’s most significant revelations. “Leaders are eager to adopt these new AI tools, but they don’t trust AI to touch live environments,” he said. “The lack of trust is shown in the data; 98% have lower trust in AI operating in production than in coding assistants.”

The observability industry built for human-speed engineering is falling short in the age of AI

The findings raise pointed questions about the current generation of observability tools from major vendors like Datadog, Dynatrace, and Splunk. Seventy-seven percent of the engineering leaders surveyed reported low or no confidence that their current observability stack provides enough information to support autonomous root cause analysis or automated incident remediation.

Maimon did not shy away from naming the structural problem. “Major vendors often build ‘closed-garden’ ecosystems where their AI SREs can only reason over data collected by their own proprietary agents,” he said. “In a modern enterprise, teams typically have a multi-tool stack to provide full coverage. By forcing a team into a single-vendor silo, these tools create an uncomfortable dependency and a strategic liability: if the vendor’s data coverage is missing a specific layer, the AI is effectively blind to the root cause.”

The second issue, Maimon argued, is that current observability-backed AI SRE solutions offer only partial visibility — defined by what engineers thought to log at the time of deployment. Because failures rarely follow predefined paths, autonomous root cause analysis using only these tools will frequently miss the key diagnostic evidence. “To move toward true autonomous remediation,” he said, “the industry must shift toward AI SRE without vendor lock-in; AI SREs must be an active participant that can connect across the entire stack and interrogate live code to capture the ground truth of a failure as it happens.”

When asked what it would take to trust AI SREs, the survey’s respondents coalesced unanimously around live runtime visibility. Fifty-eight percent said they need the ability to provide “evidence traces” of variables at the point of failure, and 42% cited the ability to verify a suggested fix before it actually deploys. No respondents selected the ability to ingest multiple log sources or provide better natural language explanations — suggesting that engineering leaders do not want AI that talks better, but AI that can see better.

The question is no longer whether to use AI for coding — it’s whether anyone can trust what it produces

The survey was administered by Global Surveyz Research, an independent firm, and drew responses from Directors, VPs, and C-level executives in SRE and DevOps roles at enterprises with 1,500 or more employees across the finance, technology, and information technology sectors. Responses were collected during January and February 2026, with questions randomized to prevent order bias.

Lightrun, which is backed by $110 million in funding from Accel and Insight Partners and counts AT&T, Citi, Microsoft, Salesforce, and UnitedHealth Group among its enterprise clients, has a clear commercial interest in the problem the report describes: the company sells a runtime observability platform designed to give AI agents and human engineers real-time visibility into live code execution. Its AI SRE product uses a Model Context Protocol connection to generate live diagnostic evidence at the point of failure without requiring redeployment. That commercial interest does not diminish the survey’s findings, which align closely with independent research from Google DORA and the real-world evidence of the Amazon outages.

Taken together, they describe an industry confronting an uncomfortable paradox. AI has solved the slowest part of building software — writing the code — only to reveal that writing was never the hard part. The hard part was always knowing whether it works. And on that question, the engineers closest to the problem are not optimistic.

“If the live visibility gap is not closed, then teams are really just compounding instability through their adoption of AI,” Maimon said. “Organizations that don’t bridge this gap will find themselves stuck with long redeploy loops, to solve ever more complex challenges. They will lose their competitive speed to the very AI tools that were meant to provide it.”

The machines learned to write the code. Nobody taught them to watch it run.