http://ipkitten.blogspot.com/2023/12/optis-v-apple-trial-e-part-i-findings.html
This
is Trial E, one of the six related trials between Optis and Apple. Trials A – D
are technical trials about the validity and/or essentiality of certain patents
(the Judge in Trial E observed that the sample in these technical trials was
too small to allow him to draw any conclusion as to the quality of Optis’ Portfolio;
also, that the decisions of very experienced first instance judges varied on
appeal, showing that the outcomes of questions of essentiality and validity are
nuanced and extremely difficult to predict: [188]). Trial F concerned the
issues of whether Apple was permanently disentitled from relying on Optis’ ETSI undertaking
and thus may be subject to an unqualified injunction and some of Apple’s
defences to this; the appeal from it will be heard by the UK Supreme Court. Trial E determined in particular the terms of a FRAND licence and the
conduct of the parties in the negotiations. The Judgment, [2023] EWHC 1095 (Ch),
is not available from the BAILII website, but Managing IP uploaded a copy here.
This
Kat’s notes on the Judgment has two parts: the second part will concern the
FRAND Question and this first part contains everything else.
Findings
on specific issues and topics
Data
regarding the stack
Data
regarding the stack could not be obtained directly from ETSI. Apple relied on
data from Innography and Optis, from PA Consulting ([111] – [112], after
abandoning all the evidence of a witness, Ms Dwyer). But the Judge considered
them to be problematic: he was presented with no one from either PA Consulting
or Innography to speak to this data, and there was a significant amount of
evidence suggesting that no third party study was of any particular help ([114]
– [115]).
Innography
data were broadly speaking reliable, partly because Optis did not adduce any
evidence to challenge the reliability of the Innography database ([133]; see
also [132]).
PA
Consulting reviewed only essentiality and not validity ([118]). Their
conclusions were enormously unreliable, given the mass of patent families in
play and given the difficulty of the exercise ([122]). Moreover, validity should
also have been taken into account, because a qualitative approach to the stack
and to portions within the stack only makes sense if both essentiality and
validity are factored in ([123], i). The Judge observed that the number of SEPs
declared to a standard is more reliable than the more qualitative assessments
which seek to exclude from the stack declared SEPs that are invalid or not
essential ([136], i). Further, PA Consulting filtered for essentiality which was
a judgmental exercise. Making a judgement about the levels of essentiality in
the stack is unreliable and unsafe ([136], iii).
The
Judge accordingly concluded that the PA Consulting/Optis approach to determining
the stack size was not reliable ([137]).
The
quality of the Portfolio
Judging
from the evidence, validity was not really considered in the selection process
the Optis Portfolio underwent ([150], iii). Because there is no effective way
of testing the essentiality, validity, and importance of each patent in a portfolio,
having a lot of declared patents overall makes logical sense ([158], iv). Optis
asserted privilege on some documents that would confirm or gainsay some of the
assertions as to the Portfolio quality, which means that these assertions could
not be tested in cross-examination ([161]). There was no material on which the Judge
could conclude that Optis’ Portfolio was anything other than average [162]).
The
Judge saw no assurance that the essentiality rates put forward by Optis at
trial and in licencing negotiations were defensible ([169]).
Based
on the evidence the Judge noted that the Portfolio was assembled in the
following way: some litigation grade patents were included to make good any
claim in the courts on validity or essentiality; some to read onto
specifications to enrich the quality or apparent quality of the pool; and some makeweights
([177]).
The
conclusion on the quality of the Optis Portfolio was that there was no credible
evidence to prove that the Portfolio was above average in strength, either
regarding essentiality or validity ([182]).
General
approach to negotiating licences: Optis
Before
the hand-down of Unwired Planet (First Instance) [2017] EWHC 2988 (Pat) Optis’
rates were excessive ([193]). Optis had an unwavering commitment to ad
valorem rates ([199]). They tended not to record discussions and to conduct
business orally and without documentation. They relied to a considerable degree
various rates tied to and sought to leverage Unwired Planet (First Instance)
Judgment and jurisdiction. And the Birss rate in that Judgment (a combined
royalty rate of 0.6% for the PanOptis Portfolio) was further leveraged by
favourable comparisons between the quality of the portfolio before Birss J and
the quality of the Optis Portfolio ([200], i – iii). The Judge noted that Optis’
assertion on the quality of its portfolio would have had very limited effect on
a sophisticated counterparty, but it would have applied pressure against a
smaller one ([200], iv).
General
approach to negotiating licences: Apple
Apple
claimed to follow a FRAND framework (called ‘Apple Framework’ in the Judgment).
It was based on the following two starting premises (para 22 cited in [201]):
i) The smallest saleable patent
practising unit (SSPPU) for cellular SEPs is the baseband chipset, where the functioning
of the cellular standards is substantially practiced or embodied; and
ii) The proportion of the overall
cellular declared SEP royalty profits that is allocated to the cellular SEP
holders is determined to be no more than the profits of the baseband chipset.
The
Apple Framework includes two steps to arrive at the licence rate for a
portfolio (paras 27 – 28 cited in [201]).
Apple’s
methodology involves attributing a price to the overall stack. Their approach
was based on the smallest saleable patent practising unit (SSPPU). Apple argued
that the baseband chipset, the SSPPU, contained the cellular functionality of
the standards, and that this should frame the debate on the value. As an
example, if the chip was valued at US$ 25 and the cost of production was US$ 20,
the manufacturer’s profit is US$ 5. Apple argued that the baseband chipset
manufacturer should pay for the licence out of the US$ 5 profit, and that this
constituted the absolute limit payable by anyone ([212] – [217]). The Judge
rejected this argument. He considered that there was no reason why the baseband
chipset manufacturer would have to fund the licence fee to the stack out of the
US$ 5 profit. It is much more likely that the chipset manufacturers would
increase the price of their product to reflect the added value to the purchasers.
It is wrong to assume that the chipset manufacturer would absorb the costs of
the licence and not pass them on ([218]).
Apple
defended throughout the trial the importance of a licensee’s right to insist on
a patent-by-patent review of any patent in an SEP portfolio ([206]), but this
approach was not fixed but was regarded as an negotiating ploy ([207], ii). However,
it cannot seriously be suggested that Apple failed to engage or failed to take
the request for a licence seriously ([210]).
Apple
insisted on a lump sum payment.
Whether
the value attributable to the standards should be excluded?
The
Judge did not accept Apple’s proposition that the SEP owner should be rewarded for
the technology inherent in the SEP and not for any contribution the SEP made to
the standard itself: when the SEP owner enters into the FRAND undertaking not
to seek an injunction, this is a contribution going beyond the value of the SEP
itself ([228]).
The
comparable licences
Optis
and Apple each produced 19 licences, but because of the heavy redaction little
more can be discerned. The Judge noted in this part that to expect either party
to concede, ex ante and in negotiations, that a given patent or a proportion
of a portfolio was valid and/or essential would be ill-advised commercially
([245]).
The
Judge held that use of comparables is inevitably a fact-driven exercise ([290]).
While the readers may remember that in InterDigital v Lenovo [2023] EWHC
539 (Pat) Mellor J adopted an exclusionary approach with comparables,
disregarding most and eventually relied on a single prior licence LG 2017 to
derive all the rates in that Judgment, Marcus Smith J differed from that approach
and considered that at least in this case, the comparables only have value if
an inclusive approach is taken. Exclusions must be explained ([291]).
The
Judge rejected the unpacking of both valuation experts, fundamentally because
the nature of the comparables in this case has taken both far outside the zone
of their proper expertise ([312]). There is an insurmountable difficulty with
resolving three subjectivities in the unpacking ([314], [301]): the different
rates at which royalties were computed (ad valorem, per unit, or lump
sum?), cross-licences (Optis is not an implementer while Apple exchange cross-licences),
future royalties and past releases. Both experts unpacked in accordance with
their principal’s respective stances, not independently ([315]).
The
(unconcluded) negotiations between Optis and Apple
[323]
– [352] reviewed the negotiation history. On Optis, the Judge concluded that their
approach to negotiation was inept, and that they were simply bad negotiators
rather than playing ‘hardball’. Their offers were no more than a series of
demands for money, not underpinned by any particular effort at persuading the purchaser
that the price represented proper value ([354]). The Judge doubted whether even
Optis knew what its thinking was. There was a great amount of incoherence ([354],
iv). To the extent their offers could be understood ‘they were all over the
place’ ([354], v). Apple did not on the evidence seek to game the system or put
pressure on Optis through negotiating gamesmanship ([354], vii).
Regarding
Apple, the Judge did not consider that Apple failed to engage or operated in
bad faith ([357]). Apple would have departed from their Framework only very
reluctantly ([359]). The Judge commented that some aspects of the Apple Framework
were not pragmatically possible when seeking to agree terms for a portfolio of
any size, such as trying to consider the technical merits of the individual SEPs
or dealing with the patents on a patent-by-patent basis ([360]).
The
Judgment in Unwired
Planet (First Instance)
The
Judge rejected the submission that Unwired Planet (First Instance) might
properly inform his own determination of what is a FRAND rate. The principle in
Hollington v Hewthorn ([1943] KB 587, Court of Appeal), interpreted by
Hoffmann J is that ‘[i]n principle the judgment, verdict or award of another
tribunal is not admissible evidence to prove a fact in issue or a fact relevant
to the issue in other proceedings between different parties.’ ([363]). Although
the Judge perceived no divergence in law between his approach and that
in Unwired Planet (First Instance), he held that it would be
impermissible for him to consider a factual finding in that Judgment as binding
on him: Apple was not party to that litigation, and the evidence he has heard here
was not the evidence in that litigation ([364], i, iii).
Was
there hold up, hold out, and/or an abuse of dominance?
Hold
out and hold up is intrinsic to any hard commercial negotiation for a licence
([366]). The Judge accepted that the contractually enforceable obligation against
an SEP owner eliminated hold up in negotiations, subject to two qualifications:
parties should in no way be discouraged from seeking to leverage their legal
rights in commercial negotiations; and as the law in this area was only now
being settled the risk of hold up being threatened effectively may not be
discounted in the period up to the commencement of these proceedings ([368],
i). But losing the threat of hold up does not mean that a significant control
over the implementer’s ability to hold out has been removed: the threat of hold
up remained in the period up to the commencement of these proceedings, and
probably there is not such correlation between hold up and hold out ([368],
ii).
The
Supreme Court in Unwired Planet did not find that hold up or hold out
was unlawful ([369]). They are undesirable ([370]). The absence of hold up and
particularly hold out is likely to be an indicator of improper market power on
one side ([372]).
The
UK jurisdiction has articulated a legal regime which resolves the problems of hold
up and hold out, as described in [1] – [24] of this Judgment. The combination
of the FRAND promise on the part of an SEP owner and these courts’ reluctance to
impose an injunction for infringing an SEP when an implementer expresses their
willingness to enter into a licence on FRAND terms remove any serious prospect
of hold up ([377]). The answer to hold out is that the FRAND licence whose
terms the Judge sets does not incentivise hold out, by ensuring that past
infringements are fully paid for and that interest on the payments accrues
([379]).
Answering
the submission of the parties, Apple appears to have claimed that Optis’ conduct
prevented meaningful negotiations, but this was not considered even to come
close to being abusive, and even assuming some kind of abuse by Optis the Judge
did not see what loss Apple has suffered: the courts can award costs ([385]). Further,
Optis is not a dominant undertaking either: by giving its promise to ETSI Optis
cannot obtain an injunction for any declared SEP where the counterparty has
expressed willingness to obtain a licence on FRAND terms ([386] – [387]).
Turning
to Optis’ contentions on this point, it appeared to the Judge that any argument
about hold out that goes beyond the terms of the FRAND licence needs first to
assert that the hold out crossed the line from being legitimate to illegitimate,
but the Judge much doubted that such a distinction exists as a matter of law: courts
should not without good reason be drawn into policing the negotiations between
entities in a market, and he knew of no cause of action that is constituted by
an allegation of illegitimate hold out ([391]). These contentions are unfounded
and legally irrelevant ([392]).
Content reproduced from The IPKat as permitted under the Creative Commons Licence (UK).