Tempus AI Q2 2026: a $5.6M GAAP profit on a $75.9M operating loss

Tempus AI reported the June quarter after Thursday’s close, and the headline line in its own release was GAAP net income.

It is true. Net income was $5.6 million, or $0.03 per basic and diluted share, against a $42.8 million loss a year earlier and a $125.9 million loss in the March quarter (our arithmetic: the $120.3 million six-month loss less Q2’s profit). It is also, on its own, close to meaningless — because the operating loss in the same quarter was $75.9 million, and on the company’s own non-GAAP basis the quarter was a $7.7 million net loss, or −$0.04 a share.

Three profit numbers from the same release pointing in different directions is the place to start.

Scorecard

	Q2 2026	Note
Revenue	$382.5M	+22% YoY
— Diagnostics	$289.3M	+20%
— Data and applications	$93.2M	+28%
Operating loss (GAAP)	−$75.9M	non-GAAP: −$2.7M
Adjusted EBITDA	$8.0M	vs −$5.6M; +$13.6M YoY (our arithmetic)
GAAP net income	$5.6M	vs −$42.8M
GAAP diluted EPS	$0.03	vs −$0.25
Non-GAAP net loss per share	−$0.04	vs a −$0.08 analyst estimate cited by Investing.com
Cash and marketable securities	$820.7M

Segment revenues sum exactly to the reported total. The −$0.08 estimate is the figure Investing.com compared with the non-GAAP −$0.04 in its coverage of the call; it is not a company figure — see the method note.

Full-year 2026 guidance was raised to $1.595–1.605 billion, roughly 25% growth, with adjusted EBITDA of about $65 million.

Did Tempus AI issue guidance for Q2 2026?

No — Tempus does not issue quarterly revenue or earnings guidance; its outlook is full-year only. In its May 5, 2026 8-K it guided 2026 revenue to $1.59–1.60 billion and adjusted EBITDA to about $65 million; with the Q2 report it raised the revenue range to $1.595–1.605 billion and kept adjusted EBITDA at about $65 million. Against the year-ago quarter, revenue rose 22% to $382.5 million from $314.6 million, and net income was $5.6 million against a net loss of $42.8 million.

Why the three profit numbers disagree

The gap between a $75.9 million operating loss and $5.6 million of net income is about $81.5 million (our arithmetic), and all of it comes from below the operating line. The income statement spells out the bridge:

Line	Q2 2026
Loss from operations	−$75.9M
Interest income	+$3.9M
Interest expense	−$10.3M
Loss on debt extinguishment	−$11.6M
Other income, net	+$102.8M
Income taxes	−$0.3M
Losses from equity method investments	−$2.9M
Net income	+$5.6M

The single line that does the work is other income, net of $102.8 million. The release says net income “included … $98.5 million in unrealized gains on marketable securities” — a mark-to-market gain on equity Tempus holds, not money earned selling tests or licensing data. Strip that line out and the quarter is a loss.

That is exactly what the company’s non-GAAP net loss does. It removes fair-value changes on marketable equity securities, stock compensation ($55.6 million including related payroll taxes), acquisition costs, intangible amortization, equity-method losses, the debt-extinguishment loss and a few smaller items, and lands at a $7.7 million net loss, or −$0.04 per share. GAAP EPS of +$0.03 and non-GAAP EPS of −$0.04 are not in conflict; they answer different questions — the first includes a market gain on investments, the second excludes it.

The practical reading for a shareholder: a GAAP profit that rests on a $98.5 million unrealized gain will not repeat on its own. The six-month figures show it: even with $66.2 million of gains on marketable equity securities, the first half was a GAAP net loss of $120.3 million.

This is the same pattern that ran through the mega-cap reports the same week, at one-hundredth the scale. Amazon’s $53.4 billion Anthropic revaluation turned a modest operating beat into a three-fold headline beat. Microsoft booked $0.40 of its $4.74 EPS from investment marks. In each case the operating business and the reported profit told different stories. Tempus is the small-cap version.

The number that actually improved in the operations is adjusted EBITDA: $8.0 million against a $5.6 million loss a year ago, a $13.6 million swing (our arithmetic). Non-GAAP loss from operations narrowed to $2.7 million from $17.0 million. That is real operating progress.

What the revenue guide implies for the second half

Revenue grew 22% in the quarter. Full-year guidance implies about 25% growth, on 2025 revenue of $1,271.8 million (Form 10-K, filed 2026-02-24).

First-half revenue was $730.6 million, up about 28% (our arithmetic, from $570.4 million); reaching the $1.595–1.605 billion range needs about $864–874 million in the second half (our arithmetic). Against second-half 2025 revenue of about $701.4 million (our arithmetic, $1,271.8 million less $570.4 million), that is second-half growth of roughly 23–25% (our arithmetic) — slower than the first half’s 28% and close to the 22% Q2 just delivered. The revenue guide does not require an acceleration; the demanding part of the guide is EBITDA, below.

The release points to the bookings supporting that: roughly $200 million in total bookings in the quarter, including deals with BioNTech, Daiichi Sankyo, Level Set Bio and Incyte. Data and applications, the higher-margin half, grew 28% versus 20% for diagnostics — the right direction for mix.

The EBITDA guide is heavily back-loaded

This is the part worth doing arithmetic on.

Adjusted EBITDA was $8.0 million in Q2, a 2.1% margin on revenue (our arithmetic). First-half adjusted EBITDA was $5.2 million. Full-year guidance is about $65 million, or roughly 4.1% on $1.6 billion of revenue (our arithmetic).

So the second half has to deliver about $60 million (our arithmetic) — more than 90% of the year’s adjusted EBITDA, in two quarters, at a company that has just cleared breakeven on that measure. It is achievable in a business with operating leverage and rising data-licensing mix. It is also the single most likely place for the story to slip, and it is where the October report should be judged.

On the call, as transcribed by Investing.com, management said the business is on track to “achieve positive free cash flow by year-end.” For a company at this stage that milestone matters more than any GAAP line, because it is the point at which growth stops requiring outside money. First-half operating cash flow was −$80.8 million.

Personalis: a ~$1.5 billion all-stock deal against $820.7 million of cash and marketable securities

On July 20, Tempus agreed to acquire Personalis, a specialist in tumor-informed minimal residual disease testing, for $16.25 per share, which the release puts at about $1.5 billion of enterprise value. The transaction is expected to close in late Q4 2026 or early 2027.

Two things follow, and neither is in the guidance.

The guide excludes it entirely. The $1.595–1.605 billion revenue and ~$65 million EBITDA figures assume no Personalis impact. Whatever the combined company looks like in 2027, this year’s numbers do not describe it.

It is paid in shares by default. The 10-Q (Note 17) says the consideration “is planned to consist entirely of the Company’s common stock,” though Tempus “may elect to pay up to 50%” in cash. So the cost to an existing shareholder is, by default, dilution rather than cash — and any cash election would draw on an $820.7 million cash-and-marketable-securities balance that also has to fund the second-half plan and carries $1.17 billion of convertible senior notes on the balance sheet. How Tempus uses that cash option matters more to a 2027 per-share outcome than anything in this quarter’s release.

Strategically the logic is clear enough: MRD testing is a large adjacent market and Tempus has commercial infrastructure to put underneath it. Financially, a company still working toward self-funding is buying something worth roughly a year of its own revenue.

The 2027 catalyst is real, and it is 2027

The FDA approved Tempus’s tumor-only xT CDx, which the release says “will migrate tissue testing to ADLT pricing.” On the call, per the Investing.com transcript, CEO Eric Lefkofsky put it at “an estimated $200 uplift in ASP, which equates to approximately $85 million on an annual basis beginning in 2027.”

That is a meaningful margin event — reimbursement uplift on existing volume is close to pure contribution. It also does nothing for 2026. Anyone building a case on this number is underwriting a story that starts more than a year out, alongside an acquisition closing into the same period.

What the numbers leave open

On the numbers, three questions stand out: a GAAP profit carried by an unrealized investment gain, a revenue guide that implies second-half growth of roughly 23–25% (our arithmetic), close to Q2’s 22%, alongside an EBITDA guide that needs about $60 million from the second half, and an acquisition that is large relative to the company and excluded from guidance.

What would change the view

At the Q3 report:

• Adjusted EBITDA against the back-loaded ~$65 million full-year path. The clearest single test of whether the operating leverage is real.
• Revenue growth against 22%, and whether it holds near the roughly 23–25% second-half pace the guide implies (our arithmetic).
• Data and applications growth against 28%. This is the higher-margin half and the reason the mix argument works.
• Free cash flow, and whether the year-end target holds. More important than any GAAP profit headline.
• How Personalis gets paid for — whether Tempus elects any of the up-to-50% cash option, and the share count once it closes.
• Other income, net — how much of the next GAAP result is again mark-to-market on marketable equity securities.
• Any ADLT pricing detail that firms up the 2027 uplift.

The read

Tempus had a decent quarter: 22% growth, a genuine $13.6 million year-over-year swing in adjusted EBITDA, strong bookings, and better mix. It also reported a GAAP profit that rests on a $98.5 million unrealized gain rather than the operating statement, guided to a second half that has to carry almost all of the year’s EBITDA, and agreed to an acquisition of roughly $1.5 billion in enterprise value, paid in stock unless it chooses cash.

None of that makes it a bad business. It does mean the GAAP-profit headline is the least useful sentence written about this release, and the free-cash-flow target at year-end is the one that will actually settle the argument.

---

Sources and method. Company figures are from Tempus’s Q2 2026 results release (Form 8-K, Exhibit 99.1, filed 2026-07-30, accession 0001193125-26-326083) and its Form 10-Q for the quarter (filed 2026-07-30, accession 0001193125-26-326090; Note 17 for the Personalis terms), and 2025 revenue of $1,271.8 million from its Form 10-K for 2025 (filed 2026-02-24, accession 0001193125-26-066961), read directly from SEC EDGAR. Segment revenues of $289.3 million and $93.2 million sum exactly to the reported $382.5 million total. The −$0.08 per-share estimate is not company data: it is the analyst estimate Investing.com compared with Tempus’s non-GAAP −$0.04 in its earnings-call coverage, and it is not leaned on here. Call remarks — the free-cash-flow target and the ADLT uplift — are quoted from the Investing.com transcript of the July 30 call. Derived figures — the net-income bridge gap, the March-quarter loss, the adjusted EBITDA margins and second-half requirements, the implied second-half revenue growth, and first-half growth — are our own arithmetic on reported numbers. Whether the quarter beat or missed is judged only against the year-ago quarter, because Tempus issued no quarterly guidance; the full-year outlook change is measured against the guidance in its Form 8-K filed 2026-05-05 (Exhibit 99.1, accession 0001193125-26-206317). This piece makes no claim about the share-price reaction.

Related, from the same week: Amazon Q2 2026, Apple Q3 FY2026, Microsoft Q4 FY2026 and Meta Q2 2026.

Tradune shows what you actually hold across every account, and whether a position has beaten a plain index fund since you bought it. Educational analysis, not investment advice — and a pre-profit, acquisition-financing company carries risks a mega-cap does not.