Anthropic has selected Accenture as its first “embedded evaluator” partner, moving from a broad commitment to independent oversight toward a concrete safety program designed to keep pace with fast-moving AI model development. The move follows an earlier call from Anthropic CEO Dario Amodei for industry-wide measures to slow the “frontier” of AI progress and create room for safeguards.

In a three-step proposal published Sept. 12, Amodei argued that AI systems can accelerate progress through recursive self-improvement—an effect that, if left unchecked, may outgrow humans’ ability to understand and control the technology. He added that safeguards should be implemented alongside development rather than as an afterthought.

Key takeaways

• Anthropic says it will work with Accenture Faculty to evaluate models, conduct red-teaming, and test safeguards as part of its embedded evaluation plan.
• Accenture is expected to provide employee-like access to evaluators, reflecting Amodei’s first step toward stronger, independent oversight.
• Anthropic and Accenture each expect to invest at least $1 billion over the next five years, according to Anthropic’s announcement.
• The project’s mechanics are still being finalized because embedded evaluation is described as a new area.
• Anthropic says it will directly fund Accenture’s work in the near term, citing the lack of an existing system for financing independent evaluation.

From Amodei’s slowdown proposal to a concrete evaluator

Amodei’s Sept. 12 proposal was framed around safety concerns that have intensified alongside rapid model iteration. He highlighted a dynamic where AI capabilities can speed up the creation of subsequent generations of AI, calling it recursive self-improvement. In his view, that process could produce outcomes that develop faster than governance and control mechanisms.

While the proposal generated attention across the AI industry, reactions were mixed. Coincidentally, OpenAI CEO Sam Altman and SpaceX CEO Elon Musk voiced positive responses to Amodei’s approach, according to social posts cited in the original reporting. Nvidia CEO Jensen Huang, however, reportedly argued that regulation of this kind was unnecessary, as noted by a CNBC profile referenced in the source material.

Against that backdrop, Anthropic’s announcement of Accenture as a first embedded evaluator is best understood as an attempt to operationalize part of Amodei’s plan. The key idea is that independent evaluation should have access patterns closer to those of internal teams—so evaluators can test systems under realistic conditions, rather than relying on limited or purely external reviews.

What “embedded evaluation” will cover

Anthropic said Accenture and its AI business unit Faculty will help “evaluate and red-team models, conduct alignment assessments and testing model safeguards.” The company also stated that the program’s implementation details are still being worked out, emphasizing that embedded evaluation is an emerging practice.

That emphasis on development matters for investors and builders because it signals that the framework is not yet standardized. For companies attempting to meet safety expectations, the lack of mature procedures can create uncertainty about what “good” evaluation looks like in practice—especially when evaluation includes alignment checks and safeguard testing.

It also highlights a practical shift: instead of treating safety testing as an isolated stage, the partnership aims to build evaluation capacity that can run alongside model development. The source material also notes that embedded evaluation was already on Anthropic’s internal roadmap, but the Accenture partnership is intended to accelerate execution.

Funding, access, and what comes next

Anthropic’s announcement places financial backing at the center of the plan. Both Anthropic and Accenture expect to invest at least $1 billion each over the next five years, according to the company’s statement. The partnership is also described as non-exclusive, with Anthropic expecting to name additional evaluators in the coming weeks.

The company further said there is currently no established funding mechanism for independent evaluation, and that long-term support may need to come from pooled resources or government sources. Still, given what Anthropic characterized as urgency, it plans to fund Accenture’s work directly in the near term.

For market participants, this structure raises an important question: will independent evaluation become a scalable, ongoing “industry function,” or will it remain dependent on a handful of well-resourced labs and contractors? Anthropic’s plan suggests it intends to push toward the former, but the acknowledgment that an existing funding system is absent indicates the field still has institutional gaps.

Accenture, via statements attributed to its leadership, positioned the work as part of building the next generation of evaluation capability. The announcement cited Accenture’s Faculty as having experience testing and evaluating models for major AI laboratories and developing complex AI systems designed to be safer by construction.

Why this matters beyond AI headlines

Even though the story is focused on AI, the underlying issue—how to build credible oversight faster than capabilities evolve—is broadly relevant to crypto markets as well. Many blockchain and decentralized systems increasingly rely on AI-assisted automation, monitoring, and tooling. When safety and evaluation frameworks are in flux, teams that integrate AI into financial or infrastructure workflows may face additional compliance and risk questions.

Anthropic’s choice to fund and operationalize embedded evaluation also signals a shift in competitive dynamics within the broader tech stack: safety is being treated less like a policy promise and more like an engineering program with measurable activities such as red-teaming, alignment assessments, and safeguard testing. Whether that becomes an industry norm will likely depend on how clearly embedded evaluation methodologies can be standardized and verified over time.

Readers should watch whether Anthropic expands the program beyond Accenture with additional evaluators, and whether the partnership publishes enough detail for outsiders to assess how “employee-like access” is implemented in practice. The largest open uncertainty is methodological: embedded evaluation is new, and the effectiveness of safeguards will depend on how the program is structured as it scales.

Artificial intelligence (AI) has been the predominant investing theme ever since Nvidia (NVDA) released that monstrous beat in its fiscal first-quarter 2024 earnings report in May 2023. Since then, the market, particularly tech companies, has been divided into two halves. The first are companies that are seen as net AI winners and saw their valuations soar. Nvidia was the flagbearer of this trade until a few months back, and its market cap soared to over $5 trillion. The Jensen Huang-led company became the world’s most valuable company in the process, something not many envisioned — at least not so soon — before AI really took off.

At the other end are companies perceived to be net losers from AI. These include software companies and IT services firms relying on “man-hour” models. Basically, these are companies whose very business model is at risk as AI models automate many tasks.

More News from Barchart

Meanwhile, “AI-pocalypse” fears have eased, and names like Accenture (ACN) and Adobe (ADBE) have rebounded from their 2026 lows. However, Adobe has fallen nearly 15% from the highs hit earlier this month. In my previous article, I noted that it would be prudent to take profits off the table after the sharp rally in ADBE stock. With the stock now coming off those highs, let’s explore why it is a “Buy” now, particularly amid the chatter around pausing advanced AI development.

Advanced AI Pause

To begin with, I believe it is highly unlikely that U.S. companies would pause advanced AI development. First, there is mutual distrust between the various companies building AI, and given how important the technology could become in the years ahead, there would always be suspicions of companies secretly working on advanced models. Moreover, U.S. AI companies are also competing with Chinese companies, which are often accused of copying U.S. technology. Think of it this way: despite various treaties, the U.S. and the former Soviet Union secretly continued to work on nuclear weapons during the Cold War. We now have a new cold war/tech war between the U.S. and China, and neither would want to lag in advanced AI buildout.

Advertisement

Story Continues

Adobe Stock Has Come Off Its Highs

Coming back to ADBE, it fell after its fiscal Q3 2026 earnings report earlier this month, despite a “beat and raise” quarter. The company beat on both the top and the bottom line and raised its revenue guidance. However, adjusted for the Q3 beat, Adobe actually lowered the annual guidance, which the company attributed to forex headwinds in the current quarter.

Digging deeper into the earnings, Adobe’s remaining performance obligations (RPOs) grew 8% year-over-year in the quarter, which was the first time since fiscal 2023 that the metric grew in single digits. Adobe’s annualized recurring revenue growth is also slowing down, and the company expects it to rise 10.2% in the current fiscal year. The company is pivoting to a freemium model and expects this to weigh on ARR growth in the short term. The ARR of the AI business is growing fast, though, and it rose 150% in fiscal Q3. However, at $650 million, it is a tiny fraction of Adobe’s consolidated ARR of $27.5 billion.

ADBE Gets a New CEO

Adobe also announced that Anil Chakravarthy would become the next CEO effective December 1. He would take the baton from long-time CEO Shantanu Narayen, who announced his departure during the fiscal Q1 earnings call. Meanwhile, Chakravarthy’s selection was perplexing for at least a section of the market. In his note, Jefferies analyst Brent Thill said that company veteran David Wadhwani, who was the President of Adobe’s global Creativity & Productivity business, was seen as “the rational choice.” Meanwhile, Wadhwani announced his departure on the same day that Adobe announced Chakravarti as the next CEO. Jefferies expects further organizational changes as “Anil reshapes the organization.”

Advertisement

Adobe has seen a lot of churn in top leadership and is currently headhunting a new CFO after Dan Durn departed the company to join Marvell Technologies (MRVL).

While Chakravarthy has yet to outline plans to steer the company amid the AI disruption, many observers believe an external hire with deeper knowledge of the AI domain would have been a better choice.

ADBE Stock Forecast

While most analysts have raised Adobe’s target price this month following the Q3 conference call, Jefferies and JPMorgan Chase lowered their respective target prices. Overall, of the 38 analysts polled by Barchart, 23 rate ADBE as a “Hold” while nine rate it as a “Buy” or equivalent. The remaining six rate Adobe as a “Sell,” and its mean target price of $269.70 is just about 7.6% higher than current prices.

Should You Buy Adobe Stock?

After the recent correction, Adobe’s valuation has come down, and it trades at around 11x the $22.64 earnings per share (EPS) analysts expect it to post next fiscal year. The price-to-earnings (P/E)-to-growth multiple has also fallen below 1x. I find the valuations reasonable here, and the risk-reward has started to look a lot more favorable versus when I last covered the stock. While “AI pause” talks might not go much further, I see ADBE as a “Buy” at these levels.

On the date of publication, Mohit Oberoi had a position in: NVDA. All information and data in this article is solely for informational purposes. This article was originally published on Barchart.com

Advertisement

In 1996, these were the S&P 500‘s (SNPINDEX: ^GSPC) largest holdings by market cap:

1. Coca-Cola: $130.6 billion

2. ExxonMobil: $121.7 billion

3. Intel: $107.6 billion

4. Microsoft: $99.4 billion

5. General Electric: $97.4 billion

6. Merck: $86.4 billion

7. International Business Machines: $73.5 billion

8. Procter & Gamble: $72.5 billion

9. Johnson & Johnson: $66.3 billion

10. Walmart: $52.2 billion

30 years later, all of these companies remain well-known names in the current economy. But it’s fair to say that they don’t at all resemble what the index looks like today. Microsoft is the only “Magnificent 7” stock still in the Top 10. Apple was in the midst of a corporate crisis until Steve Jobs eventually returned to the company. Nvidia, Amazon, Meta Platforms, Alphabet, and Tesla weren’t even publicly traded companies (or didn’t exist) back then.

Missed AI’s “Act 1”? Act 2 Could Be 15x Bigger. Most investors think they missed the AI boat because they didn’t buy Nvidia in 2005. But according to our analysts, we’re only at the end of “Act 1″—the R&D phase. “Act 2” is the global rollout. Continue »

Economies evolve over time, sometimes significantly. When investing for the long term, you can try to pick stocks that will survive and thrive. Or you can buy an ETF that tracks the economy and provides broad exposure. That’s why the Vanguard Total Stock Market ETF (NYSEMKT: VTI) is my choice for a core holding for a multidecade portfolio.

VTI doesn’t need to pick the next big winners

Since 1958, the S&P 500 has returned an average of 11% per year. But as demonstrated just now, it’s a rapidly changing group of stocks that is driving those returns over the year.

That’s what makes the Vanguard Total Stock Market ETF so compelling. It doesn’t need to try to identify and pick individual winners. It will increase its exposure to them naturally over time as the market caps of these companies grow.

If you had owned the S&P 500’s top 10 holdings in 1996 and held them for the next 30 years, you probably would have still done fairly well. But you would have missed out on the emerging tech names that grew to dominate the market and economy over that time.

Plus, the Vanguard Total Stock Market ETF owns the smaller companies that often turn into bigger companies down the road. By investing in the S&P 500, you limit yourself to just the large companies that have already established themselves. Adding small-cap and mid-caps to the mix potentially gets you in earlier on the next decade’s leaders.

That’s why I’d own this ETF and have it act as the core of my portfolio over the next several decades. Its diversification, low cost, and ability to change over time make it an ideal long-term holding.

Should you buy stock in Vanguard Morningstar Total Stock Market ETF right now?

Before you buy stock in Vanguard Morningstar Total Stock Market ETF, consider this:

The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Vanguard Morningstar Total Stock Market ETF wasn’t one of them. The 10 stocks that made the cut are built for long-term growth and could produce monster returns in the coming years.

Consider when Netflix made this list on December 17, 2004… if you invested $1,000 at the time of our recommendation, you’d have $406,141!* Or when Nvidia made this list on April 15, 2005… if you invested $1,000 at the time of our recommendation, you’d have $1,347,745!*

That performance is why people listen. With a track record of beating the S&P 500 by 4x, Stock Advisor offers a distinct advantage. Don’t miss the latest top 10 list, available with Stock Advisor, and join an investing community built for the long haul.

See the 10 stocks »

*Stock Advisor returns as of September 19, 2026.

David Dierking has positions in Apple and Vanguard Morningstar Total Stock Market ETF. The Motley Fool has positions in and recommends Alphabet, Amazon, Apple, GE Aerospace, Intel, International Business Machines, Merck, Meta Platforms, Microsoft, Nvidia, Tesla, and Walmart. The Motley Fool recommends Johnson & Johnson. The Motley Fool has a disclosure policy.

The S&P 500 Has Returned About 11% Annually Since 1958. Here’s the ETF I’d Trust for the Next 30 Years. was originally published by The Motley Fool

Grayscale’s Zcash ETF is set to undergo a 3-for-1 forward share split, according to a filing submitted with the U.S. Securities and Exchange Commission. The change is scheduled to take effect after the market close on Sept. 28, with shareholders set to receive two additional shares for every share they hold.

In a press release cited in the ETF filing, Grayscale said the forward split is designed to reduce the price per share while increasing the total number of shares outstanding in equal proportion—leaving the value of an investment unchanged in theory.

Key takeaways

• Grayscale’s Zcash ETF (ZCSH) plans a 3-for-1 forward split effective after the Sept. 28 market close.
• Shareholders receive two extra shares per held share; the filing describes the change as proportionate, not value-accretive.
• Grayscale expects the lower per-share price to improve accessibility, referencing that the unit price had become “too high.”
• The move comes as Zcash’s broader market momentum has been strong over the past year, with the token up about 2,800% in that period, according to the ETF materials.

What the 3-for-1 split means for ZCSH holders

The SEC filing describes the mechanism clearly: at the close of trading on Sept. 28, shareholders will receive two additional shares for each share they own. Grayscale’s included explanation—referenced in the filing—frames the result as a straightforward arithmetic adjustment rather than a change in underlying value.

For example, the materials illustrate that an investor holding 10 shares valued at $300 each—totaling $3,000—would own 30 shares priced at $100 each after the split, with the portfolio’s total value remaining the same.

The practical effect for investors is largely operational. Shares typically trade at a lower nominal price after such events, which can influence how the product is perceived and how easily some investors can size positions. However, the split does not inherently alter the ETF’s exposure to its underlying asset.

Why Grayscale is lowering the per-share price

Grayscale said the forward split is expected to “decrease the price per share” of the fund while increasing the share count proportionately. The ETF’s materials connect this to accessibility concerns, noting that Zcash’s token has risen sharply over the past year and that the per-unit price had become considered too high.

According to the filing-related press release, ZEC has increased by about 2,800% over the last year, and that surge helped push the ETF unit price to a level Grayscale deemed less convenient for potential investors. The split is therefore positioned as a way to make the ETF easier to buy and track in everyday terms.

For traders, these changes often matter most around implementation, including how order sizes and price targets are recalibrated. For longer-term investors, the key question is whether sentiment and liquidity improve as the share price becomes more “consumer-friendly,” even though the economics should remain proportionate.

Broader Zcash market momentum and ETF spotlight

While the share split is a structural adjustment inside the ETF wrapper, the timing also lands amid renewed attention on Zcash itself. Cointelegraph previously reported that Zcash had gained about 20% over a 24-hour period after Paradigm co-founder Matt Huang disclosed an unspecified purchase of ZEC.

In that coverage, Huang characterized Zcash as a “private complement to Bitcoin” and discussed the importance of long-term funding for the project, particularly as AI-driven cyber capabilities and quantum computing advance. Earlier coverage also linked Zcash’s performance to the visibility of privacy-focused assets in a market that remains sensitive to both regulation and technological narratives.

The Block reported that ZEC climbed as high as $1,521 early Friday, which would have been viewed as an effective new all-time high for the token before retreating slightly. That context matters for ETF holders and prospective investors: when the underlying asset experiences volatility and headline-driven flows, structural moves like share splits can draw additional attention to the vehicle, even if the split itself is not a market catalyst.

What investors should watch next

The most immediate checkpoint is the Sept. 28 record date tied to the forward split, since the share ratio will take effect after the market close. After that, investors should monitor how ZCSH trades relative to its adjusted share price—especially how liquidity and bid-ask spreads behave around the split window.

More broadly, the sustained relevance of ZCSH will likely track Zcash’s next fundamental and technical developments, alongside market demand for privacy-oriented assets. Even with the per-share price coming down, the underlying question for investors remains unchanged: whether the ETF continues to attract steady inflows as ZEC’s volatility and narrative momentum evolve.

Related references: Grayscale’s SEC filing includes details of the forward split (SEC EDGAR: zcsh-ex99_1.htm), and Cointelegraph previously reported on Zcash’s short-term price action and Matt Huang’s disclosure of a ZEC purchase (Cointelegraph coverage).

Hong Kong has jailed former China Construction Bank (Asia) relationship manager Lam Chun-yin for four years after he admitted accepting more than $470,000 in Tether to authenticate false bank instruments carrying a stated value above $1.6 billion.

The Independent Commission Against Corruption said on Sept. 18 that District Court Judge Ernest Lin Kam-hung sentenced the 32-year-old after his guilty plea to one count of conspiracy for an agent to accept advantages under Hong Kong’s Prevention of Bribery Ordinance and Crimes Ordinance.

The court ordered Lam to repay approximately HK$3.7 million to CCB (Asia), an amount equal to the bribes identified in the case. ICAC said the judge started from a six-year prison term and reduced it by one-third because Lam pleaded guilty, leaving a four-year sentence after finding no exceptional reason for another reduction.

Kalshi has filed with the SEC and CFTC on September 18 to list U.S. stock and ETF-linked perpetual futures, proposing 23-hour weekday trading and a 15.50% minimum customer margin.

The SEC filing says the contracts would have no preset expiration date and would be treated as security futures products, while the CFTC has not approved Kalshi’s proposal. The regulator’s public product database still listed Kalshi’s equity perpetual submissions as “Approval Pending (45)” on September 20.

Kalshi’s proposed Rule 14.11 sets trading from 6:00 p.m. ET on Sunday through 5:00 p.m. ET on Friday, with a daily maintenance window from 5:00 p.m. to 6:00 p.m. ET. The SEC notice says trading would stop during required regulatory halts affecting the underlying security, while Kalshi could set different hours for a particular contract when its rules permit.

The same SEC filing says the perpetual security futures would settle in cash. Funding payments would be exchanged at the daily settlement time, normally 4:00 p.m. ET when the regular U.S. equity session closes. Kalshi’s rule text sets a 0.002% deadband for small pricing differences and a maximum funding magnitude of 2.00%, while its customer-margin rule requires at least 15.50% of the current market value of each position.

Kalshi Klear would clear every perpetual security-futures transaction under the proposal, according to the SEC filing. The rule text uses a standard contract unit equal to 100 shares of the underlying security, although trading can use smaller permitted units under Kalshi’s specifications.

CFTC records now list separate Kalshi submissions for AAPL, TSLA, MSFT, NVDA, AMZN, SPY and QQQ perpetuals, among other equity-linked contracts filed September 18. Each of those records remained in “Approval Pending (45)” status when checked September 20, so none of the listed stock or fund perpetuals has CFTC clearance yet.

CFTC approval and a November 2 date remain ahead

Kalshi’s SEC filing states that the proposed rule change “will become effective on November 2, 2026, or such later date” as CFTC regulations permit. The SEC’s rulemaking page says public comments are due 21 days after publication in the Federal Register, but the agency had not posted a fixed comment deadline on its page as of September 20.

The SEC separately acknowledged Kalshi’s Form 1-N registration on September 8, allowing the CFTC-designated contract market to notice-register as a national securities exchange for the limited purpose of security futures. The September 18 filing supplies the product rules Kalshi wants to use for perpetual security futures, while CFTC approval remains a separate condition.

Kalshi’s proposed listing standards restrict which securities can serve as underliers. The SEC filing requires an estimated deliverable supply above 20 million shares, market capitalization of at least $100 billion and average daily transaction value of at least $450 million over the prior six months. A security listed for less than six months would need at least $1 billion in average daily transaction value over the prior month, while the rules separately require public float of at least 7 million shares.

The SEC filing gives the Commission another control after effectiveness. It says the SEC, after consulting with the CFTC, may summarily abrogate the rule change within 60 days of its effective date and require Kalshi to refile it under a different Exchange Act process.

Coinbase and Bitnomial filed competing stock-perpetual plans

SEC records show Coinbase Derivatives filed its own security-futures rule change on September 18, the same day as Kalshi. The CFTC database lists Coinbase’s “Single Stock Perpetual Futures Contract” as “Approval Pending (45),” while the SEC filing states that the CFTC had not yet approved Coinbase’s proposal.

As crypto.news reported in its Coinbase stock-perpetual filing coverage, Coinbase had already filed notice registrations earlier in September before submitting the product proposal. The SEC’s September 18 notice covers cash-settled futures on individual equities and exchange-traded fund shares, including perpetual single-stock futures.

Bitnomial Exchange filed a separate SEC proposal on September 18 covering security-futures listing standards, customer margin and related rules. CFTC records show 10 Bitnomial stock perpetual submissions in pending status, including AAPL, MSFT, NVDA, TSLA, AMZN, AVGO, MU, GOOGL and PLTR contracts.

Bitnomial’s SEC filing proposes a 24/5 weekly session from Sunday evening through Friday and a 15.25% minimum initial and maintenance customer-margin floor. The same filing says one standard contract represents 100 shares, funding would be calculated three times daily, and the exchange would halt a stock perpetual whenever the primary listing exchange halts the underlying security.

In related coverage, crypto.news reported that Payward planned to use Bitnomial’s regulated infrastructure for U.S. perpetual products, subject to regulatory approval. The company’s September plan named Bitnomial Exchange and Bitnomial Clearinghouse as the regulated entities handling listing, clearing and settlement.

Kalshi’s Bitcoin perp approval faces a separate court challenge

The CFTC approved Kalshi’s BTCPERP contract on May 29 after reviewing it as a futures contract tied to Bitcoin’s spot price. In its approval release, the Commission cautioned that perpetual contract design “may not be suitable for all asset classes” and encouraged exchanges to seek Commission review for perpetuals referencing assets outside the Bitcoin order.

Kalshi launched its CFTC-approved Bitcoin perpetual contract in early June. Kalshi has since listed other crypto and precious-metals perpetual products, while the September stock filings use the joint SEC-CFTC security-futures process because their underliers are securities. Kalshi Bitcoin perpetual launch coverage

A separate federal case remains pending over the CFTC’s Bitcoin perpetual decision. The docket in Chicago Mercantile Exchange Inc. v. Selig shows CME sued the CFTC and Chairman Michael Selig on June 18, arguing that the May 29 Bitcoin order and related perpetual-contract policy should be vacated. The lawsuit concerns the classification of the Bitcoin perpetual and the CFTC policy statement; it does not constitute a ruling on Kalshi’s September 18 stock-perpetual applications.

The federal docket shows the CFTC moved to dismiss the case on September 2, and no merits ruling had been issued by the latest docket update reviewed. Judge Colleen Kollar-Kotelly’s September 8 schedule gives CME until October 2 to respond, sets an October 16 reply deadline, requires discovery by November 13 and sets cross-motions for November 20, with later amicus and response deadlines in December.