Nvidia shares rose roughly 3% on Monday, September 28, 2026, after the company unveiled a new $150 billion stock buyback authorization — its largest single repurchase increase ever — pushing total buyback capacity to $235 billion. The move stood out because it came on a day the broader semiconductor sector fell 2.6%, with rivals AMD, Intel, and SK Hynix all dropping between 4.5% and 5.7%. At the same time, fresh data on Nvidia’s own equity investing spree resurfaced: the company’s portfolio now spans 13 publicly listed companies and 229 private companies, anchored by an Anthropic relationship with contract value exceeding $180 billion. Together, the two stories capture a single theme — Nvidia is simultaneously buying back its own stock and buying stakes in the AI companies that buy its chips, and investors are still deciding what that means for the stock’s next move.
Monday’s Move in Numbers
| Metric | Figure |
|---|---|
| NVDA share price (Sept 28 close, approx.) | ~$230 |
| Daily change | +3% (intraday high +3.6%) |
| Market capitalization | ~$5.55 trillion |
| New buyback authorization | $150 billion |
| Total buyback authorization (after increase) | $235 billion |
| Prior buyback authorization | $85 billion |
| Semiconductor sector move (same day) | -2.6% |
| AMD / Intel / SK Hynix decline (same day) | -4.5% to -5.7% |
| Nvidia investment portfolio size (public + private) | 13 public / 229 private companies |
| Anthropic contract value | $180 billion+ |
What Nvidia Actually Announced Monday
The headline news was a capital-return decision, not an acquisition in the traditional sense. Nvidia’s board approved an additional $150 billion in share repurchase authorization, on top of the roughly $85 billion that remained from prior authorizations, bringing the total to $235 billion. CEO Jensen Huang framed the decision around valuation, saying Nvidia’s own stock has become “too appetizing to ignore” as its forward price-to-earnings ratio has compressed to around 24 times — a level that has drifted closer to the S&P 500’s roughly 20 times multiple even as Nvidia’s earnings growth has continued to outpace the index by a wide margin. “Our cash generation gives us the capacity to invest in the technologies that advance this transformation and return capital to shareholders,” Huang said, tying the buyback directly to the company’s free cash flow generation from record data center demand.
Why a Buyback Sent the Stock Up While the Chip Sector Fell
What made Monday’s reaction notable wasn’t just the size of the buyback — it was the divergence from the rest of the sector. The broader semiconductor index fell 2.6% the same day, and chip peers AMD, Intel, and SK Hynix each dropped between 4.5% and 5.7%, suggesting a sector-wide bout of profit-taking or renewed concern about AI capital-expenditure sustainability. Nvidia’s buyback effectively worked as a company-specific counter-signal: a $150 billion commitment to repurchase shares is a direct statement from management that it views the current price as attractive, regardless of what’s happening to sentiment across the rest of the industry. Buybacks also mechanically support a stock’s price over time by reducing share count, which increases earnings per share for a fixed level of profit — a dynamic investors tend to reward immediately when the size of the program is large enough to move the needle, as $150 billion clearly is against Nvidia’s roughly $5.5 trillion market cap.
The Other Side of “Nvidia Buying Stock”: Its AI Investment Portfolio
Buybacks are Nvidia purchasing its own stock. Separately, and just as relevant to Monday’s news cycle, Nvidia has spent the past year building one of the most aggressive corporate venture portfolios in tech history — buying equity stakes in the very companies that buy its GPUs. According to data cited from CLS as of September 28, Nvidia’s investment portfolio now comprises 13 publicly listed companies and 229 private companies. Earlier reporting put the portfolio’s total value at roughly $99 billion, split across three buckets: about $47.7 billion in publicly traded stock, $47.9 billion in non-marketable securities (private company stakes), and $3.3 billion in equity-method investments where Nvidia has enough influence to use equity accounting.
The largest single positions sit with the frontier AI labs Nvidia’s chips power most directly: up to $30 billion committed to OpenAI, up to $10 billion to Anthropic, and $5 billion to Safe Superintelligence, Ilya Sutskever’s AI research lab. On the public-equity side, Nvidia holds a $5 billion stake in Intel (invested in September 2025), roughly $2 billion each in Coherent and Marvell Technology, another $2 billion in Lumentum via convertible preferred shares, and positions tied to CoreWeave, Nokia, and SoftBank’s SB Energy vehicle, among others.
Anthropic’s $180 Billion Contract and the “Circular Financing” Debate
The standout figure in Monday’s portfolio disclosure was Anthropic’s contract value with Nvidia, reported at more than $180 billion — a combination of compute purchase commitments layered on top of Nvidia’s direct equity stake in the company. That structure, where a chipmaker invests in a customer that then commits to spending heavily on that same chipmaker’s hardware, has become one of the most scrutinized dynamics in AI investing. Critics have labeled it “circular financing” or “vendor financing,” arguing it can inflate the appearance of demand: Nvidia’s investment gives Anthropic capital that flows back to Nvidia as revenue, which supports Nvidia’s stock price, which in turn supports the value of Nvidia’s stake in Anthropic. Nvidia and its backers counter that the arrangement reflects genuine, enormous compute demand rather than financial engineering — pointing to OpenAI’s own disclosed plan to spend approximately $856 billion on compute and infrastructure through 2030, much of it flowing through Nvidia contracts, as evidence that the underlying demand is real rather than manufactured. Reasonable observers can and do disagree on how much weight to put on each side of that argument, and it remains one of the more important open questions hanging over the AI capex cycle.
13 Public, 229 Private: How Big Nvidia’s Bet Has Gotten
The scale of Nvidia’s private-company exposure is easy to understate. Having stakes in 229 private companies means Nvidia is effectively running a venture capital operation layered on top of its core chip business, with exposure that ranges from small strategic checks to multi-billion-dollar commitments in frontier labs. That breadth means Nvidia’s fortunes are now tied not just to how many GPUs it can manufacture and sell, but to the success of a sprawling ecosystem of AI companies it has helped fund. If a meaningful share of that ecosystem struggles to generate revenue that justifies its own spending, Nvidia would likely feel the effects both as a chip vendor with slowing order growth and as an investor marking down its equity stakes — a dual exposure that didn’t exist to nearly the same degree even two years ago.
Is Nvidia Actually Cheap Here? The Valuation Case
Huang’s framing of the buyback rests on the idea that Nvidia’s valuation has become more reasonable relative to its growth. Nvidia’s trailing price-to-earnings ratio sits around 29 times, well below its own five-year median of roughly 56 times, and its forward P/E of about 24 times is now approaching the S&P 500’s roughly 20 times multiple despite Nvidia growing revenue far faster than the index average. One third-party valuation model (GuruFocus’s “GF Value”) pegs Nvidia’s estimated intrinsic value at roughly $398 per share against its ~$230 trading price — implying the stock trades more than 40% below that model’s fair-value estimate. That figure is one analytical framework’s output, not a consensus fact, and readers should treat any single valuation model as one data point rather than a target price. Still, directionally, a management team putting $150 billion of its own capital behind its stock is a meaningful real-money signal about how it views the current price, separate from what any outside model says.
The Growth Backdrop Supporting the Buyback
The buyback isn’t happening in a vacuum — it’s backed by genuinely strong underlying numbers. Nvidia’s most recent quarterly revenue came in at $96.2 billion, with data center revenue up 117% year-over-year, and return on equity above 70%. Huang has guided that Nvidia expects to sell twice as many chips next year and is targeting roughly 70% revenue growth through fiscal 2028, aiming for approximately $673 billion in annual revenue by that point. That guidance is underpinned by demand commitments like OpenAI’s roughly $856 billion compute and infrastructure spending plan through 2030, much of which is expected to flow through Nvidia hardware purchases, as well as broader industry financing arrangements in which Big Tech companies are reportedly backing up to $300 billion in AI data center and chip debt with corporate guarantees.
Risks and Reasons for Caution
- Circular financing concerns: Nvidia’s equity stakes in customers like OpenAI and Anthropic, paired with those same companies’ massive compute spending commitments, invite scrutiny over whether reported demand is fully organic
- Insider selling without offsetting buying: Nvidia insiders sold approximately $2.21 billion in stock over the trailing 12 months with no reported insider purchases, a pattern worth watching even as the company itself buys back shares
- Sector-wide fragility: the same-day 2.6% drop across semiconductors, and sharper 4.5-5.7% declines at AMD, Intel, and SK Hynix, suggests broader nervousness about AI capex durability that a single company’s buyback doesn’t resolve
- Concentration risk in the AI portfolio: with exposure to 229 private companies and outsized bets on a handful of frontier labs, a slowdown in AI-native revenue growth across that ecosystem could pressure both Nvidia’s chip orders and its investment marks simultaneously
- Valuation models vary widely: intrinsic-value estimates like GuruFocus’s GF Value depend heavily on the assumptions fed into them and should not be treated as price targets or guarantees
What This Means for Crypto Markets
Nvidia’s move matters to crypto traders for two reasons. First, Nvidia has become one of the most closely watched bellwethers for broader risk appetite; a $150 billion buyback landing alongside strong guidance tends to support “AI-adjacent” risk assets more broadly, including tokens tied to decentralized compute, GPU rendering, and machine-learning infrastructure narratives. Second, the same circular-financing debate playing out around Nvidia’s equity stakes in AI labs echoes concerns crypto traders already apply to token ecosystems with heavy insider or treasury involvement — it’s a useful reminder to separate genuine demand signals from capital that’s simply being recycled between related parties. Tokens like Render (RENDER), which focuses on decentralized GPU rendering and compute, and Fetch.ai (FET), an autonomous-agent and AI-infrastructure token, are among the crypto assets most frequently cited as correlated to the broader Nvidia-AI compute narrative, since their usage cases sit adjacent to the same GPU-driven demand story.
Trading AI-Themed Crypto on KCEX
Traders looking to position around the AI-compute narrative that Nvidia’s news reinforces can trade RENDER/USDT and FET/USDT on KCEX with 0% maker and taker fees on spot trading. As with any narrative-driven trade, token price action can diverge sharply from the underlying equity story, so position sizing and independent research matter as much as the theme itself.
FAQ: Nvidia’s Buyback and AI Investment Portfolio
Did Nvidia acquire a company on September 28, 2026?
No new acquisition was announced that day. The news driving the stock’s roughly 3% gain was a $150 billion increase to Nvidia’s share buyback authorization, bringing the total to $235 billion. Separately, updated data on Nvidia’s existing investment portfolio (stakes in 13 public and 229 private companies) also drew attention the same day.
Why did Nvidia’s stock rise while other chip stocks fell?
The $150 billion buyback authorization acted as a company-specific signal of confidence, while the broader semiconductor sector fell 2.6% amid what appears to be sector-wide caution about AI capital spending sustainability. AMD, Intel, and SK Hynix all declined between 4.5% and 5.7% the same day.
How large is Nvidia’s AI investment portfolio?
As of the most recent disclosures, Nvidia’s portfolio includes stakes in 13 publicly listed companies and 229 private companies, with earlier reporting valuing the overall portfolio at roughly $99 billion, including large commitments to OpenAI, Anthropic, and Safe Superintelligence.
What is the “circular financing” criticism of Nvidia’s AI investments?
It refers to concerns that Nvidia investing in AI companies that then spend heavily on Nvidia chips can create the appearance of stronger demand than exists organically, since capital effectively flows in a loop between Nvidia and its customers-turned-investments.
Is Nvidia stock actually undervalued?
One third-party model (GuruFocus’s GF Value) estimated Nvidia’s fair value at roughly $398 per share against a ~$230 trading price, implying meaningful undervaluation. That is a single model’s output based on its own assumptions, not a consensus target, and should be treated as one input among many rather than a guarantee.
This article is for informational purposes only and does not constitute financial advice. Stock and cryptocurrency prices are volatile and past performance is not indicative of future results. Figures are current as of September 28, 2026 and may change. Always conduct independent research before making investment decisions.
