Materiality Check

Two accounting students bringing daily business, finance, and accounting news and content! Not financial advice!

[Materiality Check @materialitychck]

🚨Materiality Check: The challenges of investing in Anthropic/Claude

@AnthropicAI and @claudeai have been making headlines for their new models that feature advanced co-work abilities.

You as an investor may be asking: How can I get exposure to this great company, even though they are private?

Two massive public companies have significant stakes in Anthropic: $GOOGL and $AMZN.

The problem, though, is accounting rules allow for these companies to omit disclosure of both their ownership percentage in Anthropic, and the exact carrying value of the investment on their balance sheet.

This makes it impossible for investors to know both the piece of Anthropic that they would get in a potential investment, and how much value each of the companies stands to gain upon an expected $350 billion IPO.

Amazon owns anywhere from 7.8% to 19% of Anthropic depending on how much they have been diluted and how much gain they have already recognized (this is how investors are forced to calculate ownership percentage). That's anywhere from $27-67 billion -- a massive range.

We know that Google owns 14% of Anthropic, but not because they mention it in any financial reports. Instead, because it was revealed in the legal filings of an antitrust case.

Still, per our last post on Google, they only listed roughly $38 billion in non-marketable securities in last year's annual report, and have not disclosed exactly how much gain they have recognized on each of their investments.

This makes it very hard to know how much of Anthropic's rise is priced in, which might be important for investors who are specifically looking to get a piece of the Anthropic pie.

At IPO, Google's investment would be worth an estimated $49 billion.

This is information that investor's should have easy access to -- forcing them to make complex calculations and guesswork is absurd.

These investments are material and they should be treated as such!

[Materiality Check @materialitychck]

This is great.

[litquidity @litcapital]

We call it “synthetic pre-IPO exposure”. It lets you predict the SpaceX price without owning any actual shares in the company

[Materiality Check @materialitychck]

Maybe it’s just the algorithm, but the amount of tweets I’ve seen saying that $AMZN owns 20% of Anthropic is insane. Until they disclose the actual percentage, we have no idea — it could be as low as 8%.

[Materiality Check @materialitychck]

And here’s another one…. see the problem?

[Kris Patel 🇺🇸 @KrisPatel99]

I hope the market at some point remembers that $AMZN owns 15-20% of Anthropic.

Between that and forward EBITDA of 250B in 2027… $AMZN is in the same place as $GOOGL last year.

[Materiality Check @materialitychck]

This is exactly the problem — nothing against @_masterinvestor, but this is a misleading statement. We do not know how much of Anthropic that $AMZN owns. If investors think they own 15%, but in reality, that stake was diluted to 8%, there is a real problem.

Perhaps they will share the exact percentage in their next earning’s report, but for now, we cannot make decisions based on this statement alone.

[Ian Jakovan Dunlap @_masterinvestor]

Amazon owns 15% of Anthropic

[Materiality Check @materialitychck]

@_masterinvestor We don’t know how much they own. Anywhere from 8-19%.

[Materiality Check @materialitychck]

Claude makes some great points here. We are by no means advocating for greater disclosure requirements than necessary — that’s not good for anyone.

Nor are we suggesting that the $AMZN or $GOOGL investments in Anthropic are reason enough to invest.

We are only saying that for obviously material investments (like Anthropic), investors should clearly know the ownership percentage and unrealized gains that companies have.

And investors should never be forced to do complex calculations or guesswork for material information that the company certainly has.

[John Reed Stark @JohnReedStark]

This is a terrific analysis and powerful recommendation for investing in what I view as one of the better AI tools in the world (alongside Grok and ChatGPT). So I asked Anthropic (https://t.co/GuZFtfuXEO) their opinion -- here is Claude's response (anybody care to ask Grok?):

This is an interesting piece on the disclosure challenges around investing in private companies through public company stakes. Let me share my perspective on the claims and the broader thesis.

What the post gets right:

The core argument about disclosure opacity is legitimate. Under accounting rules, companies must regularly update the reported value of their private investments to reflect changes in market price per share, through "mark-to-market" adjustments
— but they often don't have to disclose exact ownership percentages or break out specific investments from their broader portfolio of non-marketable securities. This genuinely does create challenges for investors trying to calculate exposure.

The Google 14% figure is accurate — it was previously public knowledge that Google had pumped billions into the company, but its ownership stake was only revealed through legal filings in an antitrust case.

Where the numbers need updating:

The $350 billion IPO figure is actually close to Anthropic's current private market valuation, not a hypothetical. On December 31, 2025, it was confirmed that Anthropic had signed a term sheet for a $10 billion funding round at a $350 billion valuation. Wikipedia This followed the September 2025 Series F that valued Anthropic at $183 billion.

On Amazon's stake, there's useful reverse-engineering from earnings reports: Amazon's Q3 profit included a $9.5 billion pretax gain tied to the rising valuation of its Anthropic investment GeekWire when the valuation nearly tripled to $183B, which is how analysts derive that ~7.8% estimate. But you're right — it's imprecise because you need to know prior carrying values.

My take:

The frustration about disclosure is understandable, though there are legitimate reasons companies don't break out every minority investment. That said, the post slightly overstates the investment thesis — even if you could perfectly calculate Amazon's and Google's Anthropic exposure, it would represent a relatively small percentage of those companies' overall value. Amazon's market cap is roughly $2.4 trillion, so even a $25-30 billion Anthropic stake is around 1% of the company.

I'm not in a position to recommend whether this approach to "investing in Anthropic" makes sense, John — that depends on your broader portfolio goals and risk tolerance. But the analytical points about disclosure are well-taken, and it's the kind of granular accounting critique you'd appreciate from someone like Matt Levine.

[Materiality Check @materialitychck]

@Steezehuman Hello! Check us out if you like accounting!

[Materiality Check @materialitychck]

@cryptorover @grok 0, from an accounting perspective

[Materiality Check @materialitychck]

Claude for Excel is extremely powerful— however, I feel like this makes the need for high-skilled workers even more extreme. Sure, Claude may be able to make an Excel model better than someone who has average skill, but Claude will make mistakes. Companies need people of high-skill to catch these mistakes, otherwise they will be over-reliant on AI and inevitably deliver inaccurate work.

[Materiality Check @materialitychck]

@The_AI_Investor It’s a cyclical business. The question is whether or not AI has broken that cycle. Just did a longer post on it!

[Materiality Check @materialitychck]

Micron Technology, $MU, has had a generational run due to the memory shortage. They're projections for next quarter are astounding:

- $18.7 billion in revenue (133% YoY growth)
- 68% gross margin (almost doubling YoY)
- $8.2 diluted earnings per share (480% YoY growth)

This comes as they are at maximum capacity in 2026 due to data center demand, and Micron is therefore spending on new manufacturing facilities that will likely begin output in 2027.

Seems like a marvel from an accounting perspective, and certainly undervalued with a 12.5 forward P/E.

But here's the thing: is memory still cyclical?

In the past, periods of high demand in memory has been followed by periods of oversupply. This causes memory companies like Micron to suffer because they've spent tons to expand manufacturing capacity to not ultimately reach the demand they thought.

This is what could happen to them again with the $200 billion they're investing right now.

However, some think that AI has broken this cycle, and that data centers will create perpetual demand that outweighs the massive cost of these facilities.

Regardless of which way you think it goes, AI has brought a completely new scale to the memory market, and it will always have the need for HBM.

If the former is true, Micron's margins will be crushed as they will not be able to charge as much for memory. It's just a matter of when.

If the latter is true, Micron may only be at the beginning.

[John Reed Stark @JohnReedStark]

Would buying Google stock make sense as a play on SpaceX, Anthropic and Waymo? Are Google’s investment holdings already baked into the stock price even though Google’s investment holdings are not specifically referenced in Google’s SEC filings (which is seemingly permitted under SEC regulations and is consistent with Generally Accepted Accounting Principles)?

In other words, Google, with a mere forward PE ratio of 29.33 and with great products and great management , seems poised to experience future investment windfalls of epic proportions. Are investors missing the fact that Google might be the next great stock of the decade?

We already own some Google, is it a good time to buy more?

(Not investment advice; not any sort of investment professional; just an old SEC guy who likes talking stocks every once in a while.)

[Materiality Check @materialitychck]

You can't undervalue $GOOGL's amazing ability as a venture capital fund. SpaceX, Anthropic, Waymo, and this list goes on... shows incredible management, decision-making, and feel for the future. I just hope we get to see some of these gains itemized in the financial statements!

[John Reed Stark @JohnReedStark]

@materialitychck This is possibly a hidden windfall for anyone who invests in Google. Crazy that the stock price might not have this fully priced in because of accounting rules. Google’s $900 million stake will be worth $112 billion. For context, Google’s latest annual income was $98 billion!

[Materiality Check @materialitychck]

🚨Materiality Check: $GOOGL's investment in SpaceX!

SpaceX is set to IPO this year, and it could be at a valuation of up to $1.5 trillion. In 2015, $GOOGL invested $900 million for ~7.5% of SpaceX - that would be worth $111 billion at IPO.

$GOOGL did not mention SpaceX once in their most recent 10-K. The investment would presumably be included in $GOOGL's non-marketable securities account on their balance sheet, but the entire account totaled just over $37 billion in 2024, and $GOOGL does not specifically disclose what other private investments are in this account.

Unrealized gains related to SpaceX would likely be classified under "other income" on the income statement. $GOOGL recognized a ~$3.7 billion unrealized gain on investments in 2024.

How will we see these account balances change both when $GOOGL reports 2025 earnings on February 4th, and after SpaceX goes public?

$GOOGL's net income in 2024 was just over $100 billion, so this is certainly material!