Over the course of the previous eighteen months, the topic of AI investment has changed from “which companies will benefit eventually” to “which ones are already printing money and how long can it last.” The first stock to make that question feel urgent was Nvidia, which gave a loud response with returns that caused the majority of Wall Street to rush to catch up.
However, Nvidia’s remarkable performance has also raised a secondary question that many investors are now debating in more private moments: where is the real opportunity if you missed Nvidia or are searching for the next leg of the trade? Comparing Broadcom and Palantir says more about the current status of AI investing than nearly anything else. Both companies have emerged as serious answers to that question.
| Broadcom (AVGO) — Key Metrics | |
| 1-Year Stock Return | +95% — outpacing Palantir over the past twelve months despite being a significantly larger company by market capitalisation |
|---|---|
| Q1 FY2026 AI Revenue | $8.4 billion — up 106% year over year, driven by surging demand for custom AI accelerators and AI networking infrastructure |
| Total Q1 FY2026 Revenue | $19.3 billion — a record quarter, up 29% year over year; Q2 FY2026 guidance set at approximately $22 billion |
| EBITDA Margin | 68% of revenues — adjusted EBITDA of $13.1 billion in Q1; free cash flow of $8.01 billion representing 41% of revenues |
| Return on Equity (ROE) | 47.5% — meaningfully higher than Palantir, reflecting stronger capital efficiency across the business |
| Forward P/E Ratio | 31.73 — considered reasonable relative to growth rate; substantially lower than Palantir’s valuation multiple |
| Palantir (PLTR) — Key Metrics | |
| 1-Year Stock Return | +47% — strong absolute return but trailing Broadcom over the same period; driven by AIP adoption across government and commercial clients |
| Q4 2025 Total Revenue | $1.4 billion — up 70% year over year; U.S. commercial segment up 137% YoY to $507 million; government segment up 66% YoY to $570 million |
| Full-Year 2026 Guidance | $7.182–$7.198 billion — up from $4.475 billion in 2025; remaining deal value from U.S. commercial clients reached $4.38 billion, up 145% YoY |
| Free Cash Flow Margin | 56% in Q4 2025 — exceptionally strong; GAAP net income of $609 million representing a 43% margin |
| Forward P/E Ratio | 99.55 — significantly elevated; Michael Burry has cautioned the stock may be in bubble territory, with potential competition from enterprise AI rivals including Anthropic |
| Key Risk | Heavy reliance on U.S. government spending — revenue exposed to policy shifts, budget delays, and potential contract reductions under changing administration priorities |
If you look at the numbers, Broadcom has been the cleaner tale. In its fiscal first quarter of 2026, the company’s AI revenues hit $8.4 billion, up 106% year over year. This growth rate would be remarkable for a startup, let alone one with the size and complexity of Broadcom. Custom AI accelerators and AI networking infrastructure—the kind of pick-and-shovels hardware that big tech companies require to construct and run the data centers that run their AI models—are the specific and structural need driving that figure.
Although Broadcom doesn’t produce the high-profile processors that Nvidia does, it does produce a significant portion of the components that connect and transfer data surrounding those chips. The company has been expanding more quickly than even optimists anticipated, and the margins—EBITDA at 68% of revenues—indicate that it has actual pricing power in addition to revenue growth.
In light of this, Broadcom’s one-year stock return of 95% is remarkable. This is not a fast-growing, sentiment-driven business. This is a huge, well-established semiconductor and infrastructure company that, by being truly helpful in a real boom, has increased its market value by about a trillion dollars in just one year. With free cash flow of $8.01 billion in a single quarter, or 41% of revenues, the business is able to maintain investing without using capital markets as a source of funding. In the current AI moment, where most of the excitement is well ahead of the underlying economics, the Broadcom story has a steadiness that is uncommon.
Palantir is a more complex case, and rather than covering it up with enthusiasm, it is important to be open about that complexity. The growth figures are quite remarkable: revenues in Q4 2025 increased 70% year over year, the U.S. commercial segment increased 137%, and the remaining deal value from commercial clients increased 145%. These are not ambiguous or soft numbers. The free cash flow margin of 56% in the most recent quarter is the kind of figure that serious investors notice, and the Artificial Intelligence Platform, or AIP, is obviously finding genuine demand in both government and business contexts. Palantir is not a company that poses as a business. It operates a business.
The valuation at which all of this is priced is the issue, and it is a serious one. With a forward price-to-earnings ratio of 99.55, investors are paying a significant premium for Palantir’s potential future growth, leaving virtually little room for disappointment. When investor Michael Burry, who gained notoriety for correctly predicting the 2008 mortgage crisis, openly warns that a stock might be in bubble territory, it’s worth at least taking a step back.
In the data analytics and government AI arena, where Palantir has previously had no direct competition, Burry has explicitly highlighted the challenge posed by enterprise AI firms like Anthropic. Although it’s still uncertain if such competitive challenges will materialize in a significant way in the near future, the stock’s capacity to withstand negative surprises is limited with a P/E of about 100.
The concentration of government revenue is another factor to take into account. Because government contracts, both domestic and foreign, account for almost half of Palantir’s revenue, the company is exposed to political and budgetary factors that are not present in purely commercial enterprises. Contract awards may be postponed. Priorities in the budget can change. No amount of product excellence can completely counteract changes in the revenue forecast caused by an administration change or a review of defense budget.
When comparing these two companies side by side, it seems that Palantir is the higher-conviction wager for investors who sincerely think the company’s platforms will continue to compound at extraordinary rates, while Broadcom is the stock for investors who want exposure to the AI buildout without having to pay bubble-era prices for it. Both points of view can be justified. Only one of them needs everything to work out.