Our Take
Novartis is buying optionality on Antares' discovery engine, not a proven pipeline — the lead candidate is still preclinical, and Novartis controls which targets get licensed.
Why it matters
Oncology is Novartis' $16.8 billion revenue engine (2025), and undruggable targets represent the next frontier. The deal structure (capped upfront payment, milestone-driven licensing) signals confidence in Antares' platform but risk-shifts to the biotech.
Do this week
Oncology investors: map which Antares targets Novartis is likely to license by studying the 'limited' set disclosed in the collaboration agreement when it becomes public.
The deal structure and timeline
Novartis agreed to pay $105 million upfront to Antares Therapeutics, plus up to $1.8 billion in milestone payments tied to exercise, development, regulatory, and commercial events, plus royalties. Antares will lead all research on an undisclosed but "limited" number of oncology targets until Novartis decides to license specific assets.
Antares' lead precision oncology program is expected to enter clinical trials in 2026. The company currently has multiple additional candidates in preclinical development. Antares spun out from Scorpion Therapeutics in June 2025 with three preclinical candidates and $177 million in Series A funding. Scorpion itself exited earlier that year when Eli Lilly paid up to $2.5 billion for its PI3Kα pipeline (January 2025).
Antares says it tackles undruggable targets using proprietary compound libraries, mass spectrometry experimentation and computation, and novel pocket-finding approaches. Adam Friedman, Antares CEO, framed the partnership as allowing the company to "scale" its discovery engine with Novartis' development and commercial infrastructure.
Oncology is Novartis' growth engine, and nothing is undruggable yet
Novartis oncology sales grew 18 percent to $16.8 billion in 2025 (company-reported), making it the Swiss drugmaker's largest therapeutic segment. Breast cancer drug Kisqali generated $4.8 billion alone. In that context, a $1.9 billion bet on a discovery platform for conventionally intractable targets makes strategic sense.
But the timing and structure matter. Antares has no marketed drugs and no Phase 1 readout yet. Novartis is not buying a pipeline; it is renting a discovery method. The company retains the right to cherry-pick which targets to license, leaving Antares to cover R&D costs upfront. Novartis' recent deal activity (spending up to $3 billion on Synnovation's PI3Kα inhibitors in March, and $2 billion on Excellergy in May, per the source) suggests the company is willing to buy late-stage assets outright. This Antares arrangement looks like a platform bet.
What to watch and what to do
Novartis' disclosure of the "limited" number of targets will be the key. If Antares' Series A backers believed the company could address 50 or more undruggable oncology targets, Novartis' restriction to a subset signals the company sees productivity risk or capacity constraints in Antares' engine.
The milestone schedule will also reveal how much Novartis believes in speed. If the clinical trial readout in 2026 triggers immediate exercise decisions, Antares has real momentum. If Novartis stretches the evaluation phase, the biotech is exposed to funding risk and long dilution timelines.
Investors in early-stage oncology discovery platforms should audit whether their own models match Novartis' apparent skepticism about platform breadth. If Novartis sees 3 to 5 viable targets, not 50, that is the market signal.